🌿 Leahhha

W10 - My Final Project Proposal

Dec 02, 20245 min read

My Final Project Proposal

Transforming Memories and Images into Art 🌐

Concept

The project explores how large language models (LLMs) can act as creative companions, helping artists reinterpret memories and images into unique forms of art. By tapping into LLMs’ capability to generate abstract or surreal interpretations, artists can experiment with new ways of transforming everyday visuals into complex, abstract, or “crazy” artwork, with options for users to explore travel suggestions inspired by their image submissions.

Aims

  1. Transform Memories into Art
    Reimagine and reinterpret personal memories through unique artwork generated from uploaded images, turning familiar moments into fresh, visually engaging art.

    • Target Audience:
      • Everyone: Whether users are casual art lovers or just looking for a fun experience, they can enjoy seeing how AI transforms their photos into artistic images and poetic descriptions.

  2. Generate Abstract Ideas
    Allow users to see everyday moments from a new angle, creating abstract art or poetic interpretations inspired by the contents of their images.

    • Target Audience:
      • Artists and Creatives: Uploading their previous works, drawings, or concepts, artists can explore new ideas through visual and textual reinterpretations that add a fresh perspective to their original concepts. These AI-driven abstract ideas may further inspire and invigorate their creative process.

  3. Suggest New Places to Travel
    Recognize and analyze visual themes within user-uploaded images to suggest travel destinations with similar aesthetic qualities, offering users a travel experience in sync with their visual tastes.

    • Target Audience:
      • Artists and Inspiration Seekers: For those seeking inspiration or simply looking for new travel ideas, this feature provides personalized recommendations of places that resonate with the visual essence of their uploaded images.

This project is designed to speed up creative workflows, serve as a second brain for ideation, and make art generation and travel inspiration an interactive experience.

Features & Workflow

1. Image Recognition & Interpretation

  • Input: Users upload an image (e.g., a photo from a memorable trip, an object, or a place).
  • Process: The image recognition model identifies and tags objects, colors, and themes present in the photo.
  • Output: Keywords and descriptive words that summarize the image content (e.g., “coastal beach sunset” or “dense forest with misty hues”).

2. Image to Art & Poetry Generator

  • Visual Artwork Generation:
    • Input: Recognized keywords from the image.
    • Process: Use AI (linking with Midjourney or similar APIs) to generate abstract or realistic art based on keywords.
    • Output: New images inspired by the photo but with artistic reinterpretation, which could be abstract, surreal, or stylized as per user preferences.
  • Poem Generation:
    • Input: Keywords derived from image recognition.
    • Process: Use NLP to generate a short poem, haiku, or prose that conveys the mood or theme of the image.
    • Output: A poetic representation of the scene or objects in the photo, reflecting the emotions or setting of the image.

3. Image to Travel Suggestions

  • Process: Based on keywords, the website would match the photo with similar geographical features or landmarks around the world.
  • Output: Recommended travel destinations that match the aesthetic or natural features in the image. For example, if the image contains keywords like “mountainous,” “alpine lake,” the website might suggest locations like Lake Tahoe, the Swiss Alps, or Banff.

4. Webpage Layout & User Experience

  • Image Input Section: A section where users can upload their images and see the breakdown of keywords generated by the recognition model.
  • Art & Poetry Display: Once processed, the page would showcase the generated art and poem alongside the original image, letting users experience their memory through a new lens.
  • Travel Recommendation Section: Displays travel destinations, with clickable links to explore more about each suggested place.

Tools & Technologies

  • Frontend: React or Vue.js to create an interactive and responsive UI.
  • Backend: Node.js or Django to handle file uploads, process image recognition, and communicate with external APIs.
  • Image Recognition API: Google Vision API, Amazon Rekognition, or similar services for object and scene recognition.
  • Art Generation: Integration with Midjourney or DALL-E for visual artwork.
  • Text Generation: OpenAI GPT model or a similar NLP model to create poetic interpretations based on keywords.

Future Potential

This project could expand to support a gallery feature where users can save their favorite artworks and poems. Additionally, it could include social sharing options for users to share their memories re-imagined as art and poetry, creating a community of shared experiences and inspirations.

In class note

Domain Name: a person’s name rather than numbers

input output: txt2img img2img

stability.ai: generative images (probably the most open source one)

AI tools from red square talk:

  1. Pika: Cool effect |300

Allowing artist to participate with AI tool and how can it leads us The artistic tone: flatten effect on images is pretty (how will it affect people’s esthetic taste? how should we, as audiences, view this type of generative art)

It is hard for generative film to capture human’s facial expression It is definetly good creating environment

  1. florafauna: AI tool can, real time video connection with touch designer.

Going through class documentation

Artistic Works

  • 🌿 Leaves of Manifold by Helena Sarin
    • It is also autumn season.
    • Leaves = someone’s life; no two leafs are a like so is people. = like puzzles
    • The color on those autumn leafs are indeed very inspiring and reviving

Graph View

Backlinks

  • No backlinks found